We faced very strange situation, after some time (approx every few hours) CAPSMan on CCR1036 drops all APs one by one and then registering them from scratch (empty → selecting channels → running ap)
Controller CCR1036, APs count 120 ( not all on caps, in sum should be 300 APs)
Software version on CCR updated to 38.1 (before was 36 which was quite buggy and CAPSMan was freezing without update of policy)
Software on APs - 70% already on 38.1, another’s 36
Can it be some L2/L3 loop ?
P.S. concurrent users ~ 600, CPU load on this time 1-3%
Thanks for your help, it’s very important for us to prolong using your devices in next plans of expansion
I’ve started seeing this under heavy load. Also seeing large amounts (about equal to wan) traffic on the bridge interface that is layer2 traffic. Seems to behave itself quite well on the ccr1016 till it’s 50mbit+ of wan traffic. See massive CPU spikes and caps drop off due to timeouts. Started removing capsman on these deployments.